Introduction
Hemoperitoneum, the accumulation of free blood in the abdominal cavity, is a common reason for dogs to present to the hospital on an emergency basis. Hemoperitoneum can be the result of trauma (or coagulopathy following trauma) but is frequently seen as a consequence of spontaneous tumor rupture, with the spleen being the organ most commonly affected.1–5 For dogs diagnosed with splenic tumor rupture, emergency surgical intervention is typically required to remove the spleen and associated tumor and to control the ongoing blood loss. Splenectomy for most dogs with benign splenic tumor rupture is often curative. This stands in stark contrast to the beginning of a complex treatment path for dogs with malignant splenic tumor rupture or the choice to halt further diagnostic or therapeutic efforts due to a guarded prognosis. Given the emergent presentation of these cases, the decision to pursue surgery needs to be made by the pet owner without knowing whether the underlying tumor is due to benign or malignant disease, with a final histologic diagnosis not being available until several days after operation.
The use of abdominal ultrasonography is often included in the care of these patients. The purpose of this ultrasonography is both to determine the origin of the bleeding tumor and, more importantly, to look for the presence of possible metastasis within the abdomen. Specifically, the presence of liver or other intra-abdominal nodules on a preoperative abdominal ultrasonography may reasonably raise concerns for possible metastatic disease and may dissuade some owners from electing to go forward with treatment. Previous research investigating the prevalence of hepatic metastasis secondary to splenic hemangiosarcoma at the time of diagnosis is limited. Clendaniel et al6 observed an overall 37% prevalence of hepatic metastasis from splenic hemangiosarcoma, consistent with findings by Hammond and Pesillo-Crosby.7 Another study5 reported that the presence of nodules on both the spleen and liver indicated a 48% likelihood of a splenic hemangiosarcoma diagnosis. A retrospective study8 of 14 dogs with retroperitoneal sarcoma reported an overall prevalence of metastasis to be 29%. Nine of the dogs in that study had a diagnosis of retroperitoneal hemangiosarcoma, with 3 of those dogs having concurrent metastasis, resulting in a hemangiosarcoma-specific prevalence of 33%.
Recently we reported9 on a small cohort of dogs with splenic tumor rupture and found that only 33% of liver nodules found on preoperative ultrasonography were subsequently histologically confirmed to be metastatic hemangiosarcoma lesions. Accordingly, these preliminary data raise questions about the value of ultrasonography as a preoperative staging diagnostic for identifying potential metastatic disease in the liver. Furthermore, ultrasound failed to identify most omental metastases that were subsequently identified intraoperatively in this same cohort.
Based on these collective data, we were prompted to inquire whether relying solely on conventional abdominal ultrasonography as a method to detect the bleeding source and potential visceral metastasis is satisfactory for deciding to proceed with surgery. To address this query, we examined data obtained from a comprehensive prospective study (Ethos Precision Medicine Umbrella Study for Hemangiosarcoma [Ethos-PUSH]) involving dogs experiencing hemoperitoneum attributed to suspected splenic tumor rupture.
Methods
Case selection criteria
This study included dogs with nontraumatic hemoperitoneum secondary to splenic tumor rupture that were being enrolled in a nationwide, prospective, multicenter randomized trial evaluating novel treatments for hemangiosarcoma (Ethos-PUSH). Dogs were enrolled through 20 emergency and specialty hospitals across the US from October 2020 to May 2022. The inclusion criteria required dogs to have a diagnosis of nontraumatic hemoperitoneum secondary to a ruptured splenic tumor, have no evidence of pulmonary metastasis on preoperative thoracic radiographs, and be scheduled to undergo splenectomy surgery with a board-certified surgeon or emergency department doctor.
A full abdominal ultrasonography was required for all dogs, which was performed by a board-certified radiologist or a board-certified internist; however, abnormal findings did not define eligibility of patients for the study. Abnormal splenic and liver lesions were defined as any discrete and measurable mass or nodule.
Post hoc analysis of Ethos-PUSH data
From the prospectively enrolled Ethos-PUSH cohort, dogs were eligible for inclusion in our study if they had preoperative abdominal ultrasonographic findings and intraoperative findings reported in a case report form or within the electronic medical record. Furthermore, each dog was required to have any abnormal lesions sampled for histopathology during operation (including those identified on preoperative ultrasonography). The ultrasonographic, surgical, and histopathologic results were recorded in a central database.
Statistical analysis
Statistical analysis was performed with a statistical software program (Prism, version 10.1.2; GraphPad Software Inc). Normality was assessed with Shapiro-Wilk tests, and descriptive data were reported with medians and ranges. Fisher exact tests were used to compare the presence of preoperative liver lesions to splenic tumor diagnosis and liver lesion diagnosis. A significance level of α = .05 was used.
Results
Patient demographics
Ninety-nine dogs (36 female dogs [35 spayed and 1 intact] and 63 male dogs [62 neutered and 1 intact]) met the eligibility criteria for this study (Table 1). The median age of the dogs included was 10 years (range, 3 to 13 years; Table 2). The study group consisted of 29 mixed-breed dogs, 14 German Shepherd Dogs, 9 Golden Retrievers, 9 Labrador Retrievers, 5 Australian Shepherds, 4 Boxers, 3 French Bulldogs, 3 Goldendoodles, 3 Labradoodles, 3 Pembroke Welsh Corgis, 2 Great Pyrenees, 2 Siberian Huskies, and 1 each of Belgian Malinois, Bernese Mountain Dog, Boerboel, Border Collie, Chesapeake Bay Retriever, Chinese Crested, Cocker Spaniel, Dachshund, Doberman Pinscher, Greater Swiss Mountain Dog, Italian Greyhound, Newfoundland, and Portuguese Water Dog. There were similar breed demographics between dogs with splenic hemangiosarcoma and those with benign disease. The dogs had a median body weight of 30 kg (range, 7.1 to 58 kg).
Sex demographics of study population.
| Entire population (n = 99) | Splenic HSA (n = 58) | Benign splenic lesions (n = 33) | |
|---|---|---|---|
| Intact female | 1 (1%) | 1 | 0 |
| Spayed female | 35 (35%) | 19 | 14 |
| Intact male | 1 (1%) | 0 | 0 |
| Neutered male | 62 (63%) | 38 | 19 |
HSA = Hemangiosarcoma.
Age and weight demographics of study population.
| Entire population (n = 99) | Splenic HSA (n = 58) | Benign splenic lesions (n = 33) | |
|---|---|---|---|
| Median age (range) | 10 y (3–13 y) | 10 y (5–13 y) | 10 y (3–13 y) |
| Median weight (range) | 30 kg (7.1–58 kg) | 30.6 kg (7.6–58 kg) | 24.8 kg (7.1–51.3 kg) |
Concordance of liver lesion identification
Twenty of 99 (20%) dogs had liver lesions identified on their preoperative ultrasonography. An additional 22 dogs had liver lesions identified intraoperatively that were not seen on preoperative ultrasonography. There was no correlation between nodule identification on preoperative ultrasonographic and intraoperative findings (Table 3). There was a sensitivity of 40%, specificity of 72%, positive predictive value of 27%, and negative predictive value of 83%. When extrapolated to the whole cohort, 22 of 99 (22%) dogs did not have their liver lesions identified on preoperative ultrasonography.
Correlation between preoperative ultrasonographic and intraoperative liver findings.
| Dogs with preoperative US nodules (%) | Dogs without preoperative US nodules (%) | |
|---|---|---|
| Intraoperative nodules | 8 (40) | 22 (28) |
| No intraoperative nodules | 12 (60) | 57 (72) |
| P value | .29 |
US = Ultrasonographic.
Association of liver lesion identification to splenic diagnosis
In this cohort, 58 of 99 (59%) dogs were diagnosed with splenic hemangiosarcoma, 33 (33%) dogs were diagnosed with a benign splenic lesion, and 8 (8%) were diagnosed with other splenic malignancies (4 with stromal sarcoma, 2 with histiocytic sarcoma, 1 with carcinoma, and 1 with leiomyosarcoma). Seventeen of the 20 (85%) dogs with liver lesions identified on preoperative ultrasonography were diagnosed with benign splenic lesions (including 9 with nodular hyperplasia, 4 with hepatocellular glycogen-type vacuolation, and 4 with hematopoietic hyperplastic nodules). The 3 (15%) remaining dogs received malignant diagnoses (2 with hemangiosarcoma and 1 with carcinoma). There was a significant association between the presence of liver lesions on preoperative ultrasonography and splenic tumor diagnosis, with the presence of liver lesions being associated with a higher likelihood of a benign splenic tumor diagnosis (P < .0001).
Association of liver lesion identification to histologic diagnosis
Of the 20 dogs with liver lesions identified on preoperative ultrasonography, 16 of these had a liver biopsy collected during operation. Liver biopsies were not collected in the remaining 4 dogs due to the absence of visual liver lesions intraoperatively. There were 49 dogs that did not have ultrasonographic liver lesions identified but still had a liver biopsy collected during operation due to the identification of lesions intraoperatively (29 focal-to-diffuse mottled hepatic parenchyma, 13 nodules, and 7 masses). Of the 16 dogs with liver nodules on preoperative ultrasonography, only 3 (19%) of these dogs were diagnosed as metastasis (2 with hemangiosarcoma and 1 with carcinoma), with the remaining 13 (81%) dogs receiving benign diagnoses (8 with nodular hyperplasia, 4 with hepatocellular glycogen-type vacuolation, and 1 with hematopoietic hyperplastic nodule). The lesions reported on the ultrasonography for the 3 dogs with malignant tumors were identified as 1 to a few small hypoechoic nodules measuring between 2 and 4 mm. The lesions reported on the ultrasonography as benign were identified as follows: 3 had diffuse mottled parenchyma with hyperechoic nodules measuring more than 5 mm, 4 had 1 to a few hyperechoic nodules measuring more than 10 mm, and 10 had hypoechoic nodules measuring more than 5 mm. Of the 49 dogs with no liver nodules seen on preoperative ultrasonography, 9 (18%) were diagnosed with metastatic disease and 40 (82%) were diagnosed with benign lesions. There was no association between the identification of liver lesions on preoperative abdominal ultrasonography and the presence of metastatic disease on liver biopsy (Table 4). There was a sensitivity of 19%, specificity of 82%, positive predictive value of 25%, and negative predictive value of 75%.
Correlation between preoperative ultrasonographic findings and metastatic disease diagnosis.
| Dogs with preoperative US nodules (%) | Dogs without preoperative US nodules (%) | |
|---|---|---|
| Metastatic disease | 3 (19) | 9 (18) |
| Nonmetastatic disease | 13 (81) | 40 (82) |
| P value | > .99 |
Identification of lesions in other organs
Eleven of the full cohort of 99 (11%) dogs had lesions identified in organs other than the spleen and liver on their preoperative ultrasonography. These included 2 lesions involving the mesentery and 1 each involving the omentum, cecum, adrenal gland, kidney, and mesenteric lymph node. There were also 4 lesions that weren’t noted to be associated with a specific organ. Only 3 (27%) of these lesions were identified intraoperatively (2 in the omentum and 1 in the cecum).
Seventeen of 99 (17%) dogs were identified with lesions in organs other than the spleen and liver during operation. The lesions noted intraoperatively included 13 in the omentum and 1 each in the stomach, cecum, pancreas, and diaphragm. Fourteen (82%) of these lesions were missed on the preoperative ultrasonography.
Fifteen of the 17 dogs with intraoperative lesions not associated with the spleen or liver had a biopsy collected from the organ during operation. This included 12 from the omentum and 1 each from the pancreas, cecum, and a splenic lymph node. Of the omental nodule biopsies, 6 of 12 (50%) were diagnosed as malignant hemangiosarcoma and the remaining 6 of 12 (50%) were classified as hematomas. The pancreatic nodule was identified as inflammation consistent with pancreatitis, the cecal lesion was identified as carcinoma, and the lymph node was identified as reactive. Of the 3 lesions reported on preoperative ultrasonography that were identified intraoperatively, 1 of them was found to be true metastatic carcinomatosis (33%).
Discussion
Preoperative ultrasonography is frequently utilized to assess for potential metastasis to inform surgical planning. However, to date, there remains a lack of evidence regarding the association between the identification of lesions on preoperative ultrasonography and the diagnosis of malignant or metastatic disease. For instance, in a study conducted by Millar and Zersen,10 the presence of cavitated splenic masses on abdominal ultrasonography did not support a diagnosis of malignancy. Furthermore, relying solely on lesion identification via ultrasonography may not offer a reliable means of determining the presence of metastatic disease. This concern is underscored by the findings of our study, wherein only 4 out of 31 (13%) dogs with preoperative ultrasound-detected lesions suggestive of possible metastasis (20 in the liver and 11 in other organs) were confirmed to have true metastatic disease.
An important limitation of ultrasonography for preoperative staging is that the entire abdomen may not be able to be imaged due to the size/conformation of the dog, the presence of bowel gas, obesity, or anatomical barriers such as the rib cage. Hemangiosarcoma can also present as microscopic lesions or involve multiple sites, some of which may not be clearly discernible on ultrasonography. One study11 highlighted significant variability in the ultrasonographic appearance of hepatic parenchymal lesions across all diagnoses, with no statistically significant associations observed between ultrasonographic appearance and diagnosis. In the present study, hepatic lesions on abdominal ultrasonography for dogs with metastatic disease were described as hypoechoic nodules measuring between 2 and 4 mm. However, this same appearance of hepatic lesions was also observed in most of the benign lesion cases; therefore, there were no discernable differences in ultrasonographic appearance to differentiate between benign and malignant lesions. With the new and growing field of artificial-intelligence–guided diagnostic imaging interpretation, the ability to differentiate these lesions may become possible in the future. In the study by Millar and Zersen,10 the presence of cavitary masses or nodules did not support a diagnosis of malignancy; however, the study did suggest that abdominal ultrasonography may not be helpful in distinguishing malignant from benign lesions but could aid in surgical planning and confirming the presence of a bleeding mass. Therefore, if the trigger for deciding whether to pursue surgical intervention is based on lesions identified on preoperative ultrasonography, there are concerns that this could lead to dogs with benign disease that would be curable with surgery alone to instead be euthanized over the fear of possible metastatic disease.
In the context of liver lesions identified on preoperative ultrasonography in dogs with ruptured splenic tumors, it is essential to recognize that the presence of these lesions does not unequivocally imply liver metastasis. The prevalence of hepatic metastasis has been subject to examination across several studies. Pintar et al12 observed a notably high prevalence of 47.6%. Conversely, studies by Hammond and Pesillo-Crosby7 and Carloni at al13 reported lower prevalences ranging from 34.4% to 37%. Other studies14,15 have documented comparatively lower prevalence rates between 28% and 30%.
Our study contributes significantly to this discussion by reporting an overall prevalence of liver metastasis in dogs with hemoperitoneum secondary to a ruptured splenic tumor at a markedly lower rate of only 12%, indicating a notable departure from previously reported figures. An explanation for this difference could be the prospective nature of this study, compared to the retrospective nature of most of these previous studies, demonstrating the potential impact of underlying biases inherent within retrospective data.
Eleven of the 12 (92%) dogs diagnosed with hepatic metastasis had hemangiosarcoma and the remainder had carcinomatosis. Furthermore, we showed that only 15% of liver lesions identified on preoperative ultrasonography were histologically confirmed to be metastatic, whereas the remainder of the lesions were determined to be benign lesions. Similar findings have been noted in another study.9
Given the inherent limitations of ultrasonography as a staging diagnostic, CT scans or MRI could be more ideal approaches to patient staging. They can provide detailed visualization of affected organs and tissues, aid in surgical planning, and assist in assessing the overall prognosis. Furthermore, these modalities allow for the scanning of the whole dog and not just a single body cavity, allowing for the opportunity to detect metastatic lesions that otherwise likely would not be seen on ultrasound (ie, intramuscular metastases).10,14,16,17 The major disadvantages of CT and MRI are the cost to the owner, their limited availability outside of regular business hours, the need for trained personnel to perform the imaging, and the additional time required under anesthesia in a potentially unstable patient.
The incidence of hepatic nodular hyperplasia and other similar benign pathologies increases with age, with a previous study18 reporting up to 70% of older dogs showing the presence of benign liver lesions. It is important to recognize the prevalence of these age-related liver lesions when discussing the results of staging diagnostics with pet owners.
In the present study, 17% of patients had intraoperative lesions identified in organs other than the liver or spleen; 18% of those lesions were suspected on preoperative ultrasonography. This demonstrates a low sensitivity for metastasis identification with ultrasonography in other organs, similarly as in the liver.
A potential limitation of this study was that splenectomy procedures were performed by a combination of board-certified surgeons and emergency department doctors. All veterinarians performing the splenectomy procedures for these dogs had undergone specific training in biopsy collection for this study; however, the difference in experience level could still have introduced a bias due to missed lesions that weren’t sampled. Additionally, it cannot be definitively confirmed that the lesions identified in organs other than the liver and spleen on preoperative abdominal ultrasonography had biopsies collected from the same lesion site once in operation.
In summary, these data suggest favorable outcomes for dogs with hemoperitoneum, with a lower incidence of metastasis than historically believed. Abdominal ultrasonography, when used as a preoperative staging diagnostic, has been shown to have a low sensitivity for the detection of metastatic lesions. Therefore, a more comprehensive approach, such as whole-body CT or MRI, may be necessary and is an active area of investigation for the authors. Nonetheless, further advancements in diagnostic imaging may allow for improved detection of metastatic lesions, such as higher-resolution ultrasound machines and the use of artificial intelligence platforms.
Acknowledgments
None reported.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.
Funding
The authors have nothing to disclose.
ORCID
A. Ramirez https://orcid.org/0000-0002-4729-7927
S. Stewart https://orcid.org/0000-0003-1148-2589
References
- 1.↑
Aronsohn MG, Dubiel B, Roberts B, Powers BE. Prognosis for acute nontraumatic hemoperitoneum in the dog: a retrospective analysis of 60 cases (2003-2006). J Am Anim Hosp Assoc. 2009;45(2):72-77. doi:10.5326/0450072
- 2.
Cleveland MJ, Casale S. Incidence of malignancy and outcomes for dogs undergoing splenectomy for incidentally detected nonruptured splenic nodules or masses: 105 cases (2009-2013). J Am Vet Med Assoc. 2016;248(11):1267-1273. doi:10.2460/javma.248.11.1267
- 3.
Day MJ, Lucke VM, Pearson H. A review of pathological diagnoses made from 87 canine splenic biopsies. J Small Anim Pract. 1995;36(10):426-433. doi:10.1111/j.1748-5827.1995.tb02769.x
- 4.
Fernandez S, Lang JM, Maritato KC. Evaluation of nodular splenic lesions in 370 small-breed dogs (< 15 kg). J Am Anim Hosp Assoc. 2019;55(4):201-209. doi:10.5326/JAAHA-MS-6934
- 5.↑
Leyva FJ, Loughin CA, Dewey CW, Marino DJ, Akerman M, Lesser ML. Histopathologic characteristics of biopsies from dogs undergoing surgery with concurrent gross splenic and hepatic masses: 125 cases (2012-2016). BMC Res Notes. 2018;11(1):122. doi:10.1186/s13104-018-3220-1
- 6.↑
Clendaniel DC, Sivacolundhu RK, Sorenmo KU, et al. Association between macroscopic appearance of liver lesions and liver histology in dogs with splenic hemangiosarcoma: 79 cases (2004-2009). J Am Anim Hosp Assoc. 2014;50(4):e6-e10. doi:10.5326/JAAHA-MS-6059
- 7.↑
Hammond TN, Pesillo-Crosby SA. Prevalence of hemangiosarcoma in anemic dogs with a splenic mass and hemoperitoneum requiring a transfusion: 71 cases (2003-2005). J Am Vet Med Assoc. 2008;232(4):553-558. doi:10.2460/javma.232.4.553
- 8.↑
Liptak JM, Dernell WS, Ehrhart EJ, Rizzo SA, Rooney MB, Withrow SJ. Retroperitoneal sarcomas in dogs: 14 cases (1992-2002). J Am Vet Med Assoc. 2004;224(9):1471-1477. doi:10.2460/javma.2004.224.1471
- 9.↑
Stewart SD, Ehrhart EJ, Davies R, Khanna C. Prospective observational study of dogs with splenic mass rupture suggests potentially lower risk of malignancy and more favourable perioperative outcomes. Vet Comp Oncol. 2020;18(4):811-817. doi:10.1111/vco.12621
- 10.↑
Millar SL, Zersen KM. Diagnostic value of the ultrasonographic description of a splenic mass or nodule as cavitated in 106 dogs with nontraumatic hemoabdomen. Am J Vet Res. 2021;82(12):970-974. doi:10.2460/ajvr.21.08.0130
- 11.↑
Warren-Smith CM, Andrew S, Mantis P, Lamb CR. Lack of associations between ultrasonographic appearance of parenchymal lesions of the canine liver and histological diagnosis. J Small Anim Pract. 2012;53(3):168-173. doi:10.1111/j.1748-5827.2011.01184.x
- 12.↑
Pintar J, Breitschwerdt EB, Hardie EM, Spaulding KA. Acute nontraumatic hemoabdomen in the dog: a retrospective analysis of 39 cases (1987-2001). J Am Anim Hosp Assoc. 2003;39(6):518-522. doi:10.5326/0390518
- 13.↑
Carloni A, Terragni R, Morselli-Labate AM, et al. Prevalence, distribution, and clinical characteristics of hemangiosarcoma-associated skeletal muscle metastases in 61 dogs: a whole body computed tomographic study. J Vet Intern Med. 2019;33(2):812-819. doi:10.1111/jvim.15456
- 14.↑
Ivancić M, Long F, Seiler GS. Contrast harmonic ultrasonography of splenic masses and associated liver nodules in dogs. J Am Vet Med Assoc. 2009;234(1):88-94. doi:10.2460/javma.234.1.88
- 15.↑
Boston SE, Higginson G, Monteith G. Concurrent splenic and right atrial mass at presentation in dogs with HSA: a retrospective study. J Am Anim Hosp Assoc. 2011;47(5):336-341. doi:10.5326/JAAHA-MS-5603
- 16.↑
O’Brien RT. Improved detection of metastatic hepatic hemangiosarcoma nodules with contrast ultrasound in three dogs. Vet Radiol Ultrasound. 2007;48(2):146-148. doi:10.1111/j.1740-8261.2007.00222.x
- 17.↑
Strobel D, Krodel U, Martus P, Hahn EG, Becker D. Clinical evaluation of contrast-enhanced color Doppler sonography in the differential diagnosis of liver tumors. J Clin Ultrasound. 2000;28(1):1-13. doi:10.1002/(SICI)1097-0096(200001)28:1<1::AID-JCU1>3.0.CO;2-J
- 18.↑
Bergman JR. Nodular hyperplasia in the liver of the dog: an association with changes in the Ito cell population. Vet Pathol. 1985;22(5):427-438. doi:10.1177/030098588502200501
